function SlowBuffer (size) { this.length = size; }; var assert = require('assert'); exports.INSPECT_MAX_BYTES = 50; function toHex(n) { if (n < 16) return '0' + n.toString(16); return n.toString(16); } function utf8ToBytes(str) { var byteArray = []; for (var i = 0; i < str.length; i++) if (str.charCodeAt(i) <= 0x7F) byteArray.push(str.charCodeAt(i)); else { var h = encodeURIComponent(str.charAt(i)).substr(1).split('%'); for (var j = 0; j < h.length; j++) byteArray.push(parseInt(h[j], 16)); } return byteArray; } function asciiToBytes(str) { var byteArray = [] for (var i = 0; i < str.length; i++ ) // Node's code seems to be doing this and not & 0x7F.. byteArray.push( str.charCodeAt(i) & 0xFF ); return byteArray; } function base64ToBytes(str) { return require("base64-js").toByteArray(str); } SlowBuffer.byteLength = function (str, encoding) { switch (encoding || "utf8") { case 'hex': return str.length / 2; case 'utf8': case 'utf-8': return utf8ToBytes(str).length; case 'ascii': case 'binary': return str.length; case 'base64': return base64ToBytes(str).length; default: throw new Error('Unknown encoding'); } }; function blitBuffer(src, dst, offset, length) { var pos, i = 0; while (i < length) { if ((i+offset >= dst.length) || (i >= src.length)) break; dst[i + offset] = src[i]; i++; } return i; } SlowBuffer.prototype.utf8Write = function (string, offset, length) { var bytes, pos; return SlowBuffer._charsWritten = blitBuffer(utf8ToBytes(string), this, offset, length); }; SlowBuffer.prototype.asciiWrite = function (string, offset, length) { var bytes, pos; return SlowBuffer._charsWritten = blitBuffer(asciiToBytes(string), this, offset, length); }; SlowBuffer.prototype.binaryWrite = SlowBuffer.prototype.asciiWrite; SlowBuffer.prototype.base64Write = function (string, offset, length) { var bytes, pos; return SlowBuffer._charsWritten = blitBuffer(base64ToBytes(string), this, offset, length); }; SlowBuffer.prototype.base64Slice = function (start, end) { var bytes = Array.prototype.slice.apply(this, arguments) return require("base64-js").fromByteArray(bytes); } function decodeUtf8Char(str) { try { return decodeURIComponent(str); } catch (err) { return String.fromCharCode(0xFFFD); // UTF 8 invalid char } } SlowBuffer.prototype.utf8Slice = function () { var bytes = Array.prototype.slice.apply(this, arguments); var res = ""; var tmp = ""; var i = 0; while (i < bytes.length) { if (bytes[i] <= 0x7F) { res += decodeUtf8Char(tmp) + String.fromCharCode(bytes[i]); tmp = ""; } else tmp += "%" + bytes[i].toString(16); i++; } return res + decodeUtf8Char(tmp); } SlowBuffer.prototype.asciiSlice = function () { var bytes = Array.prototype.slice.apply(this, arguments); var ret = ""; for (var i = 0; i < bytes.length; i++) ret += String.fromCharCode(bytes[i]); return ret; } SlowBuffer.prototype.binarySlice = SlowBuffer.prototype.asciiSlice; SlowBuffer.prototype.inspect = function() { var out = [], len = this.length; for (var i = 0; i < len; i++) { out[i] = toHex(this[i]); if (i == exports.INSPECT_MAX_BYTES) { out[i + 1] = '...'; break; } } return ''; }; SlowBuffer.prototype.hexSlice = function(start, end) { var len = this.length; if (!start || start < 0) start = 0; if (!end || end < 0 || end > len) end = len; var out = ''; for (var i = start; i < end; i++) { out += toHex(this[i]); } return out; }; SlowBuffer.prototype.toString = function(encoding, start, end) { encoding = String(encoding || 'utf8').toLowerCase(); start = +start || 0; if (typeof end == 'undefined') end = this.length; // Fastpath empty strings if (+end == start) { return ''; } switch (encoding) { case 'hex': return this.hexSlice(start, end); case 'utf8': case 'utf-8': return this.utf8Slice(start, end); case 'ascii': return this.asciiSlice(start, end); case 'binary': return this.binarySlice(start, end); case 'base64': return this.base64Slice(start, end); case 'ucs2': case 'ucs-2': return this.ucs2Slice(start, end); default: throw new Error('Unknown encoding'); } }; SlowBuffer.prototype.hexWrite = function(string, offset, length) { offset = +offset || 0; var remaining = this.length - offset; if (!length) { length = remaining; } else { length = +length; if (length > remaining) { length = remaining; } } // must be an even number of digits var strLen = string.length; if (strLen % 2) { throw new Error('Invalid hex string'); } if (length > strLen / 2) { length = strLen / 2; } for (var i = 0; i < length; i++) { var byte = parseInt(string.substr(i * 2, 2), 16); if (isNaN(byte)) throw new Error('Invalid hex string'); this[offset + i] = byte; } SlowBuffer._charsWritten = i * 2; return i; }; SlowBuffer.prototype.write = function(string, offset, length, encoding) { // Support both (string, offset, length, encoding) // and the legacy (string, encoding, offset, length) if (isFinite(offset)) { if (!isFinite(length)) { encoding = length; length = undefined; } } else { // legacy var swap = encoding; encoding = offset; offset = length; length = swap; } offset = +offset || 0; var remaining = this.length - offset; if (!length) { length = remaining; } else { length = +length; if (length > remaining) { length = remaining; } } encoding = String(encoding || 'utf8').toLowerCase(); switch (encoding) { case 'hex': return this.hexWrite(string, offset, length); case 'utf8': case 'utf-8': return this.utf8Write(string, offset, length); case 'ascii': return this.asciiWrite(string, offset, length); case 'binary': return this.binaryWrite(string, offset, length); case 'base64': return this.base64Write(string, offset, length); case 'ucs2': case 'ucs-2': return this.ucs2Write(string, offset, length); default: throw new Error('Unknown encoding'); } }; // slice(start, end) SlowBuffer.prototype.slice = function(start, end) { if (end === undefined) end = this.length; if (end > this.length) { throw new Error('oob'); } if (start > end) { throw new Error('oob'); } return new Buffer(this, end - start, +start); }; SlowBuffer.prototype.copy = function(target, targetstart, sourcestart, sourceend) { var temp = []; for (var i=sourcestart; i this.length) { throw new Error('oob'); } if (start > end) { throw new Error('oob'); } for (var i = start; i < end; i++) { this[i] = value; } } function coerce(length) { // Coerce length to a number (possibly NaN), round up // in case it's fractional (e.g. 123.456) then do a // double negate to coerce a NaN to 0. Easy, right? length = ~~Math.ceil(+length); return length < 0 ? 0 : length; } // Buffer function Buffer(subject, encoding, offset) { if (!(this instanceof Buffer)) { return new Buffer(subject, encoding, offset); } var type; // Are we slicing? if (typeof offset === 'number') { this.length = coerce(encoding); this.parent = subject; this.offset = offset; } else { // Find the length switch (type = typeof subject) { case 'number': this.length = coerce(subject); break; case 'string': this.length = Buffer.byteLength(subject, encoding); break; case 'object': // Assume object is an array this.length = coerce(subject.length); break; default: throw new Error('First argument needs to be a number, ' + 'array or string.'); } if (this.length > Buffer.poolSize) { // Big buffer, just alloc one. this.parent = new SlowBuffer(this.length); this.offset = 0; } else { // Small buffer. if (!pool || pool.length - pool.used < this.length) allocPool(); this.parent = pool; this.offset = pool.used; pool.used += this.length; } // Treat array-ish objects as a byte array. if (isArrayIsh(subject)) { for (var i = 0; i < this.length; i++) { if (subject instanceof Buffer) { this.parent[i + this.offset] = subject.readUInt8(i); } else { this.parent[i + this.offset] = subject[i]; } } } else if (type == 'string') { // We are a string this.length = this.write(subject, 0, encoding); } } } function isArrayIsh(subject) { return Array.isArray(subject) || Buffer.isBuffer(subject) || subject && typeof subject === 'object' && typeof subject.length === 'number'; } exports.SlowBuffer = SlowBuffer; exports.Buffer = Buffer; Buffer.poolSize = 8 * 1024; var pool; function allocPool() { pool = new SlowBuffer(Buffer.poolSize); pool.used = 0; } // Static methods Buffer.isBuffer = function isBuffer(b) { return b instanceof Buffer || b instanceof SlowBuffer; }; Buffer.concat = function (list, totalLength) { if (!Array.isArray(list)) { throw new Error("Usage: Buffer.concat(list, [totalLength])\n \ list should be an Array."); } if (list.length === 0) { return new Buffer(0); } else if (list.length === 1) { return list[0]; } if (typeof totalLength !== 'number') { totalLength = 0; for (var i = 0; i < list.length; i++) { var buf = list[i]; totalLength += buf.length; } } var buffer = new Buffer(totalLength); var pos = 0; for (var i = 0; i < list.length; i++) { var buf = list[i]; buf.copy(buffer, pos); pos += buf.length; } return buffer; }; // Inspect Buffer.prototype.inspect = function inspect() { var out = [], len = this.length; for (var i = 0; i < len; i++) { out[i] = toHex(this.parent[i + this.offset]); if (i == exports.INSPECT_MAX_BYTES) { out[i + 1] = '...'; break; } } return ''; }; Buffer.prototype.get = function get(i) { if (i < 0 || i >= this.length) throw new Error('oob'); return this.parent[this.offset + i]; }; Buffer.prototype.set = function set(i, v) { if (i < 0 || i >= this.length) throw new Error('oob'); return this.parent[this.offset + i] = v; }; // write(string, offset = 0, length = buffer.length-offset, encoding = 'utf8') Buffer.prototype.write = function(string, offset, length, encoding) { // Support both (string, offset, length, encoding) // and the legacy (string, encoding, offset, length) if (isFinite(offset)) { if (!isFinite(length)) { encoding = length; length = undefined; } } else { // legacy var swap = encoding; encoding = offset; offset = length; length = swap; } offset = +offset || 0; var remaining = this.length - offset; if (!length) { length = remaining; } else { length = +length; if (length > remaining) { length = remaining; } } encoding = String(encoding || 'utf8').toLowerCase(); var ret; switch (encoding) { case 'hex': ret = this.parent.hexWrite(string, this.offset + offset, length); break; case 'utf8': case 'utf-8': ret = this.parent.utf8Write(string, this.offset + offset, length); break; case 'ascii': ret = this.parent.asciiWrite(string, this.offset + offset, length); break; case 'binary': ret = this.parent.binaryWrite(string, this.offset + offset, length); break; case 'base64': // Warning: maxLength not taken into account in base64Write ret = this.parent.base64Write(string, this.offset + offset, length); break; case 'ucs2': case 'ucs-2': ret = this.parent.ucs2Write(string, this.offset + offset, length); break; default: throw new Error('Unknown encoding'); } Buffer._charsWritten = SlowBuffer._charsWritten; return ret; }; // toString(encoding, start=0, end=buffer.length) Buffer.prototype.toString = function(encoding, start, end) { encoding = String(encoding || 'utf8').toLowerCase(); if (typeof start == 'undefined' || start < 0) { start = 0; } else if (start > this.length) { start = this.length; } if (typeof end == 'undefined' || end > this.length) { end = this.length; } else if (end < 0) { end = 0; } start = start + this.offset; end = end + this.offset; switch (encoding) { case 'hex': return this.parent.hexSlice(start, end); case 'utf8': case 'utf-8': return this.parent.utf8Slice(start, end); case 'ascii': return this.parent.asciiSlice(start, end); case 'binary': return this.parent.binarySlice(start, end); case 'base64': return this.parent.base64Slice(start, end); case 'ucs2': case 'ucs-2': return this.parent.ucs2Slice(start, end); default: throw new Error('Unknown encoding'); } }; // byteLength Buffer.byteLength = SlowBuffer.byteLength; // fill(value, start=0, end=buffer.length) Buffer.prototype.fill = function fill(value, start, end) { value || (value = 0); start || (start = 0); end || (end = this.length); if (typeof value === 'string') { value = value.charCodeAt(0); } if (!(typeof value === 'number') || isNaN(value)) { throw new Error('value is not a number'); } if (end < start) throw new Error('end < start'); // Fill 0 bytes; we're done if (end === start) return 0; if (this.length == 0) return 0; if (start < 0 || start >= this.length) { throw new Error('start out of bounds'); } if (end < 0 || end > this.length) { throw new Error('end out of bounds'); } return this.parent.fill(value, start + this.offset, end + this.offset); }; // copy(targetBuffer, targetStart=0, sourceStart=0, sourceEnd=buffer.length) Buffer.prototype.copy = function(target, target_start, start, end) { var source = this; start || (start = 0); end || (end = this.length); target_start || (target_start = 0); if (end < start) throw new Error('sourceEnd < sourceStart'); // Copy 0 bytes; we're done if (end === start) return 0; if (target.length == 0 || source.length == 0) return 0; if (target_start < 0 || target_start >= target.length) { throw new Error('targetStart out of bounds'); } if (start < 0 || start >= source.length) { throw new Error('sourceStart out of bounds'); } if (end < 0 || end > source.length) { throw new Error('sourceEnd out of bounds'); } // Are we oob? if (end > this.length) { end = this.length; } if (target.length - target_start < end - start) { end = target.length - target_start + start; } return this.parent.copy(target.parent, target_start + target.offset, start + this.offset, end + this.offset); }; // slice(start, end) Buffer.prototype.slice = function(start, end) { if (end === undefined) end = this.length; if (end > this.length) throw new Error('oob'); if (start > end) throw new Error('oob'); return new Buffer(this.parent, end - start, +start + this.offset); }; // Legacy methods for backwards compatibility. Buffer.prototype.utf8Slice = function(start, end) { return this.toString('utf8', start, end); }; Buffer.prototype.binarySlice = function(start, end) { return this.toString('binary', start, end); }; Buffer.prototype.asciiSlice = function(start, end) { return this.toString('ascii', start, end); }; Buffer.prototype.utf8Write = function(string, offset) { return this.write(string, offset, 'utf8'); }; Buffer.prototype.binaryWrite = function(string, offset) { return this.write(string, offset, 'binary'); }; Buffer.prototype.asciiWrite = function(string, offset) { return this.write(string, offset, 'ascii'); }; Buffer.prototype.readUInt8 = function(offset, noAssert) { var buffer = this; if (!noAssert) { assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset < buffer.length, 'Trying to read beyond buffer length'); } if (offset >= buffer.length) return; return buffer.parent[buffer.offset + offset]; }; function readUInt16(buffer, offset, isBigEndian, noAssert) { var val = 0; if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 1 < buffer.length, 'Trying to read beyond buffer length'); } if (offset >= buffer.length) return 0; if (isBigEndian) { val = buffer.parent[buffer.offset + offset] << 8; if (offset + 1 < buffer.length) { val |= buffer.parent[buffer.offset + offset + 1]; } } else { val = buffer.parent[buffer.offset + offset]; if (offset + 1 < buffer.length) { val |= buffer.parent[buffer.offset + offset + 1] << 8; } } return val; } Buffer.prototype.readUInt16LE = function(offset, noAssert) { return readUInt16(this, offset, false, noAssert); }; Buffer.prototype.readUInt16BE = function(offset, noAssert) { return readUInt16(this, offset, true, noAssert); }; function readUInt32(buffer, offset, isBigEndian, noAssert) { var val = 0; if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 3 < buffer.length, 'Trying to read beyond buffer length'); } if (offset >= buffer.length) return 0; if (isBigEndian) { if (offset + 1 < buffer.length) val = buffer.parent[buffer.offset + offset + 1] << 16; if (offset + 2 < buffer.length) val |= buffer.parent[buffer.offset + offset + 2] << 8; if (offset + 3 < buffer.length) val |= buffer.parent[buffer.offset + offset + 3]; val = val + (buffer.parent[buffer.offset + offset] << 24 >>> 0); } else { if (offset + 2 < buffer.length) val = buffer.parent[buffer.offset + offset + 2] << 16; if (offset + 1 < buffer.length) val |= buffer.parent[buffer.offset + offset + 1] << 8; val |= buffer.parent[buffer.offset + offset]; if (offset + 3 < buffer.length) val = val + (buffer.parent[buffer.offset + offset + 3] << 24 >>> 0); } return val; } Buffer.prototype.readUInt32LE = function(offset, noAssert) { return readUInt32(this, offset, false, noAssert); }; Buffer.prototype.readUInt32BE = function(offset, noAssert) { return readUInt32(this, offset, true, noAssert); }; /* * Signed integer types, yay team! A reminder on how two's complement actually * works. The first bit is the signed bit, i.e. tells us whether or not the * number should be positive or negative. If the two's complement value is * positive, then we're done, as it's equivalent to the unsigned representation. * * Now if the number is positive, you're pretty much done, you can just leverage * the unsigned translations and return those. Unfortunately, negative numbers * aren't quite that straightforward. * * At first glance, one might be inclined to use the traditional formula to * translate binary numbers between the positive and negative values in two's * complement. (Though it doesn't quite work for the most negative value) * Mainly: * - invert all the bits * - add one to the result * * Of course, this doesn't quite work in Javascript. Take for example the value * of -128. This could be represented in 16 bits (big-endian) as 0xff80. But of * course, Javascript will do the following: * * > ~0xff80 * -65409 * * Whoh there, Javascript, that's not quite right. But wait, according to * Javascript that's perfectly correct. When Javascript ends up seeing the * constant 0xff80, it has no notion that it is actually a signed number. It * assumes that we've input the unsigned value 0xff80. Thus, when it does the * binary negation, it casts it into a signed value, (positive 0xff80). Then * when you perform binary negation on that, it turns it into a negative number. * * Instead, we're going to have to use the following general formula, that works * in a rather Javascript friendly way. I'm glad we don't support this kind of * weird numbering scheme in the kernel. * * (BIT-MAX - (unsigned)val + 1) * -1 * * The astute observer, may think that this doesn't make sense for 8-bit numbers * (really it isn't necessary for them). However, when you get 16-bit numbers, * you do. Let's go back to our prior example and see how this will look: * * (0xffff - 0xff80 + 1) * -1 * (0x007f + 1) * -1 * (0x0080) * -1 */ Buffer.prototype.readInt8 = function(offset, noAssert) { var buffer = this; var neg; if (!noAssert) { assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset < buffer.length, 'Trying to read beyond buffer length'); } if (offset >= buffer.length) return; neg = buffer.parent[buffer.offset + offset] & 0x80; if (!neg) { return (buffer.parent[buffer.offset + offset]); } return ((0xff - buffer.parent[buffer.offset + offset] + 1) * -1); }; function readInt16(buffer, offset, isBigEndian, noAssert) { var neg, val; if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 1 < buffer.length, 'Trying to read beyond buffer length'); } val = readUInt16(buffer, offset, isBigEndian, noAssert); neg = val & 0x8000; if (!neg) { return val; } return (0xffff - val + 1) * -1; } Buffer.prototype.readInt16LE = function(offset, noAssert) { return readInt16(this, offset, false, noAssert); }; Buffer.prototype.readInt16BE = function(offset, noAssert) { return readInt16(this, offset, true, noAssert); }; function readInt32(buffer, offset, isBigEndian, noAssert) { var neg, val; if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 3 < buffer.length, 'Trying to read beyond buffer length'); } val = readUInt32(buffer, offset, isBigEndian, noAssert); neg = val & 0x80000000; if (!neg) { return (val); } return (0xffffffff - val + 1) * -1; } Buffer.prototype.readInt32LE = function(offset, noAssert) { return readInt32(this, offset, false, noAssert); }; Buffer.prototype.readInt32BE = function(offset, noAssert) { return readInt32(this, offset, true, noAssert); }; function readFloat(buffer, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset + 3 < buffer.length, 'Trying to read beyond buffer length'); } return require('./buffer_ieee754').readIEEE754(buffer, offset, isBigEndian, 23, 4); } Buffer.prototype.readFloatLE = function(offset, noAssert) { return readFloat(this, offset, false, noAssert); }; Buffer.prototype.readFloatBE = function(offset, noAssert) { return readFloat(this, offset, true, noAssert); }; function readDouble(buffer, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset + 7 < buffer.length, 'Trying to read beyond buffer length'); } return require('./buffer_ieee754').readIEEE754(buffer, offset, isBigEndian, 52, 8); } Buffer.prototype.readDoubleLE = function(offset, noAssert) { return readDouble(this, offset, false, noAssert); }; Buffer.prototype.readDoubleBE = function(offset, noAssert) { return readDouble(this, offset, true, noAssert); }; /* * We have to make sure that the value is a valid integer. This means that it is * non-negative. It has no fractional component and that it does not exceed the * maximum allowed value. * * value The number to check for validity * * max The maximum value */ function verifuint(value, max) { assert.ok(typeof (value) == 'number', 'cannot write a non-number as a number'); assert.ok(value >= 0, 'specified a negative value for writing an unsigned value'); assert.ok(value <= max, 'value is larger than maximum value for type'); assert.ok(Math.floor(value) === value, 'value has a fractional component'); } Buffer.prototype.writeUInt8 = function(value, offset, noAssert) { var buffer = this; if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset < buffer.length, 'trying to write beyond buffer length'); verifuint(value, 0xff); } if (offset < buffer.length) { buffer.parent[buffer.offset + offset] = value; } }; function writeUInt16(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 1 < buffer.length, 'trying to write beyond buffer length'); verifuint(value, 0xffff); } for (var i = 0; i < Math.min(buffer.length - offset, 2); i++) { buffer.parent[buffer.offset + offset + i] = (value & (0xff << (8 * (isBigEndian ? 1 - i : i)))) >>> (isBigEndian ? 1 - i : i) * 8; } } Buffer.prototype.writeUInt16LE = function(value, offset, noAssert) { writeUInt16(this, value, offset, false, noAssert); }; Buffer.prototype.writeUInt16BE = function(value, offset, noAssert) { writeUInt16(this, value, offset, true, noAssert); }; function writeUInt32(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 3 < buffer.length, 'trying to write beyond buffer length'); verifuint(value, 0xffffffff); } for (var i = 0; i < Math.min(buffer.length - offset, 4); i++) { buffer.parent[buffer.offset + offset + i] = (value >>> (isBigEndian ? 3 - i : i) * 8) & 0xff; } } Buffer.prototype.writeUInt32LE = function(value, offset, noAssert) { writeUInt32(this, value, offset, false, noAssert); }; Buffer.prototype.writeUInt32BE = function(value, offset, noAssert) { writeUInt32(this, value, offset, true, noAssert); }; /* * We now move onto our friends in the signed number category. Unlike unsigned * numbers, we're going to have to worry a bit more about how we put values into * arrays. Since we are only worrying about signed 32-bit values, we're in * slightly better shape. Unfortunately, we really can't do our favorite binary * & in this system. It really seems to do the wrong thing. For example: * * > -32 & 0xff * 224 * * What's happening above is really: 0xe0 & 0xff = 0xe0. However, the results of * this aren't treated as a signed number. Ultimately a bad thing. * * What we're going to want to do is basically create the unsigned equivalent of * our representation and pass that off to the wuint* functions. To do that * we're going to do the following: * * - if the value is positive * we can pass it directly off to the equivalent wuint * - if the value is negative * we do the following computation: * mb + val + 1, where * mb is the maximum unsigned value in that byte size * val is the Javascript negative integer * * * As a concrete value, take -128. In signed 16 bits this would be 0xff80. If * you do out the computations: * * 0xffff - 128 + 1 * 0xffff - 127 * 0xff80 * * You can then encode this value as the signed version. This is really rather * hacky, but it should work and get the job done which is our goal here. */ /* * A series of checks to make sure we actually have a signed 32-bit number */ function verifsint(value, max, min) { assert.ok(typeof (value) == 'number', 'cannot write a non-number as a number'); assert.ok(value <= max, 'value larger than maximum allowed value'); assert.ok(value >= min, 'value smaller than minimum allowed value'); assert.ok(Math.floor(value) === value, 'value has a fractional component'); } function verifIEEE754(value, max, min) { assert.ok(typeof (value) == 'number', 'cannot write a non-number as a number'); assert.ok(value <= max, 'value larger than maximum allowed value'); assert.ok(value >= min, 'value smaller than minimum allowed value'); } Buffer.prototype.writeInt8 = function(value, offset, noAssert) { var buffer = this; if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset < buffer.length, 'Trying to write beyond buffer length'); verifsint(value, 0x7f, -0x80); } if (value >= 0) { buffer.writeUInt8(value, offset, noAssert); } else { buffer.writeUInt8(0xff + value + 1, offset, noAssert); } }; function writeInt16(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 1 < buffer.length, 'Trying to write beyond buffer length'); verifsint(value, 0x7fff, -0x8000); } if (value >= 0) { writeUInt16(buffer, value, offset, isBigEndian, noAssert); } else { writeUInt16(buffer, 0xffff + value + 1, offset, isBigEndian, noAssert); } } Buffer.prototype.writeInt16LE = function(value, offset, noAssert) { writeInt16(this, value, offset, false, noAssert); }; Buffer.prototype.writeInt16BE = function(value, offset, noAssert) { writeInt16(this, value, offset, true, noAssert); }; function writeInt32(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 3 < buffer.length, 'Trying to write beyond buffer length'); verifsint(value, 0x7fffffff, -0x80000000); } if (value >= 0) { writeUInt32(buffer, value, offset, isBigEndian, noAssert); } else { writeUInt32(buffer, 0xffffffff + value + 1, offset, isBigEndian, noAssert); } } Buffer.prototype.writeInt32LE = function(value, offset, noAssert) { writeInt32(this, value, offset, false, noAssert); }; Buffer.prototype.writeInt32BE = function(value, offset, noAssert) { writeInt32(this, value, offset, true, noAssert); }; function writeFloat(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 3 < buffer.length, 'Trying to write beyond buffer length'); verifIEEE754(value, 3.4028234663852886e+38, -3.4028234663852886e+38); } require('./buffer_ieee754').writeIEEE754(buffer, value, offset, isBigEndian, 23, 4); } Buffer.prototype.writeFloatLE = function(value, offset, noAssert) { writeFloat(this, value, offset, false, noAssert); }; Buffer.prototype.writeFloatBE = function(value, offset, noAssert) { writeFloat(this, value, offset, true, noAssert); }; function writeDouble(buffer, value, offset, isBigEndian, noAssert) { if (!noAssert) { assert.ok(value !== undefined && value !== null, 'missing value'); assert.ok(typeof (isBigEndian) === 'boolean', 'missing or invalid endian'); assert.ok(offset !== undefined && offset !== null, 'missing offset'); assert.ok(offset + 7 < buffer.length, 'Trying to write beyond buffer length'); verifIEEE754(value, 1.7976931348623157E+308, -1.7976931348623157E+308); } require('./buffer_ieee754').writeIEEE754(buffer, value, offset, isBigEndian, 52, 8); } Buffer.prototype.writeDoubleLE = function(value, offset, noAssert) { writeDouble(this, value, offset, false, noAssert); }; Buffer.prototype.writeDoubleBE = function(value, offset, noAssert) { writeDouble(this, value, offset, true, noAssert); }; SlowBuffer.prototype.readUInt8 = Buffer.prototype.readUInt8; SlowBuffer.prototype.readUInt16LE = Buffer.prototype.readUInt16LE; SlowBuffer.prototype.readUInt16BE = Buffer.prototype.readUInt16BE; SlowBuffer.prototype.readUInt32LE = Buffer.prototype.readUInt32LE; SlowBuffer.prototype.readUInt32BE = Buffer.prototype.readUInt32BE; SlowBuffer.prototype.readInt8 = Buffer.prototype.readInt8; SlowBuffer.prototype.readInt16LE = Buffer.prototype.readInt16LE; SlowBuffer.prototype.readInt16BE = Buffer.prototype.readInt16BE; SlowBuffer.prototype.readInt32LE = Buffer.prototype.readInt32LE; SlowBuffer.prototype.readInt32BE = Buffer.prototype.readInt32BE; SlowBuffer.prototype.readFloatLE = Buffer.prototype.readFloatLE; SlowBuffer.prototype.readFloatBE = Buffer.prototype.readFloatBE; SlowBuffer.prototype.readDoubleLE = Buffer.prototype.readDoubleLE; SlowBuffer.prototype.readDoubleBE = Buffer.prototype.readDoubleBE; SlowBuffer.prototype.writeUInt8 = Buffer.prototype.writeUInt8; SlowBuffer.prototype.writeUInt16LE = Buffer.prototype.writeUInt16LE; SlowBuffer.prototype.writeUInt16BE = Buffer.prototype.writeUInt16BE; SlowBuffer.prototype.writeUInt32LE = Buffer.prototype.writeUInt32LE; SlowBuffer.prototype.writeUInt32BE = Buffer.prototype.writeUInt32BE; SlowBuffer.prototype.writeInt8 = Buffer.prototype.writeInt8; SlowBuffer.prototype.writeInt16LE = Buffer.prototype.writeInt16LE; SlowBuffer.prototype.writeInt16BE = Buffer.prototype.writeInt16BE; SlowBuffer.prototype.writeInt32LE = Buffer.prototype.writeInt32LE; SlowBuffer.prototype.writeInt32BE = Buffer.prototype.writeInt32BE; SlowBuffer.prototype.writeFloatLE = Buffer.prototype.writeFloatLE; SlowBuffer.prototype.writeFloatBE = Buffer.prototype.writeFloatBE; SlowBuffer.prototype.writeDoubleLE = Buffer.prototype.writeDoubleLE; SlowBuffer.prototype.writeDoubleBE = Buffer.prototype.writeDoubleBE;